A whole company got
a second brain and hours back every week
Launch by Lunch built The Engine a shared, AI-powered company brain plus a set of automations tailored to how each team actually works, then taught the whole staff how to use it. Adoption stuck because it was built around the work they were already doing.
ABOUT THE ENGINE
Founded by MIT to bridge the gap between breakthrough science and real-world impact
The Engine is a nonprofit incubator and accelerator dedicated to providing early-stage Tough Tech companies with the critical infrastructure, programs, and ecosystem they need to thrive. MIT created The Engine in 2016 to support Tough Tech startups, those solving the world's most important challenges through science, engineering, and entrepreneurship, as they move from research breakthroughs to commercial impact.
Infrastructure
220,000 square feet of fully-managed wet lab, engineering, and office space, fully equipped for building Tough Tech.
Programs
Educational programming that addresses the complex challenges Tough Tech founders face, guiding them from idea to investment to impact.
Ecosystem
A thriving network of entrepreneurs, investors, academia, public officials, and corporate partners.
- Industry
- Venture & tough-tech incubation
- Team size
- 37 people
- Engagement
- Company brain + automations, now retainer
- Timeline
- April 2026 → ongoing
The Challenge
Knowledge scattered across systems, and data that was hard to keep current
The Engine works with startups and founders building hard things, supporting hundreds of them at once. That means things move fast and there are a lot of moving pieces. The knowledge that made the team effective lived across different systems or was buried in docs, so answers were hard to find and easy to lose track of.
Keeping data sources current was its own challenge: a CRM is only useful when it reflects reality, and staying that up to date in near real time took constant manual effort. On top of that, recurring work, meeting follow-ups, resident and partner updates, facilities reports, weekly status, ate real hours every week. A generic AI tool people open twice and forget was never going to solve it. They needed something built around how they actually operate.
What We Built
A two-layer brain, plus automations made for each team
One shared knowledge base for the whole company, department-specific workflows on top, and the education to make both stick.
Built the shared company brain
We stood up a single, secure, always-current knowledge base on Claude so anyone on the team can get answers without hunting down the person who knows. It refreshes on a monthly rhythm that leadership reviews in plain English, no code, no diffs, just a quick conversational sign-off.
Added a personal layer for every person
On top of the shared brain, each team member gets a private working layer the assistant remembers, their priorities, their projects, their style, so it works like a real partner, not a generic chatbot. Private to them, never shared.
Automated the recurring work, team by team
We built automations around what each team was already doing: meeting follow-ups captured and routed automatically, resident and partner reviews, weekly status, daily facilities summaries. Each one is a 'push a button, get the artifact' workflow, not one tool stretched thin across everyone.
Taught the whole company how to use it
Before and after the build, Karen and Ariel spent real time helping staff understand the foundation: what was being built, how it fit together, and why. That upfront investment in shared understanding is what turned a rollout into real adoption.
Stayed on as an ongoing partner
The Engine moved into a monthly retainer, so the brain keeps current and the team has a partner for new questions and new workflows as they grow.
The Results
Adoption stuck, and the team got time back
- Real adoption across a busy 37-person team: people picked it up quickly because it was built around work they were already doing.
- Time back for the work that matters: recurring busywork (follow-ups, status, reports) now runs automatically.
- Knowledge is findable: anyone can get answers without tracking down the one person who knows.
- Workflows that fit each team: five functional areas got automations built around how they actually operate.
- A durable foundation: the brain stays current on a monthly rhythm leadership reviews in minutes.
In Their Words
I honestly wasn't sure we'd see real adoption. Getting a busy team to change how they work is its own challenge. But people picked this up quickly because it was built around what they were already doing. Watching the team get time back in their week to focus on the work that actually moves the needle for our founders has been the real win.
The Engine exists to support companies pushing the boundaries of what's possible creatively and operationally, and how we use AI internally should reflect that. Launch by Lunch built something that fits how we actually work, not a generic tool people open twice and forget. And that makes the investment worth it.